OpenAI has recently introduced a new automatic memory management feature for ChatGPT, designed to help users manage the AI's long-term memory more easily and enhance the interactive experience. This feature is currently available only to ChatGPT Plus and ChatGPT Pro paid subscription users, and will be rolled out gradually starting today.
Previously, ChatGPT's memory function allowed users to set personalized preferences so that the model could refer to the user's background information when responding, thus providing more targeted content. For example, users could input their occupation or interests in advance, and the AI would adjust its tone and content accordingly in future conversations. However, due to memory space limits, long-term use could lead to "memory full," requiring users to manually delete old data.
The newly launched automatic memory management feature aims to solve this pain point. Once enabled, the system will automatically delete some old memories in chronological order, ensuring continuous availability of space and preventing storage overflow. Users no longer need to manage memory manually. Additionally, ChatGPT also allows users to manually adjust memory priority in the memory settings: important memories can be kept long-term, while lower-priority information will be prioritized for system cleanup.
This update makes ChatGPT's "memory" mechanism closer to how real humans manage their memories, maintaining a personalized experience while further improving usability.
